J Clin Epidemiol Vol. 52, No. 1, pp.

4955, 1999
Copyright 1999 Elsevier Science Inc. All rights reserved.

0895-4356/99/ $see front matter

PII S0895-4356(98)00139-5

Identifying Adults at Low Risk for Significant

Hyperlipidemia: A Validated Clinical Index
Steven A. Grover,1,2,3,4,5,* Carey Levinton,1 and Steeve Paquet1
1Centre

for the Analysis of Cost-Effective Care, and 2the Divisions of General Internal Medicine and
Epidemiology, The Montreal General Hospital, 4Departments of Medicine and 5Epidemiology
and Biostatistics, McGill University, Montreal, Quebec, Canada

3Clinical

ABSTRACT. The objective of this study was to develop and validate a simple clinical index to identify
individuals at increased risk of an elevated CHL/HDL ratio. Using recursive partitioning, factors associated with
an elevated CHL/HDL ratio were identified among 1993 men and 1631 women in the Lipid Research Clinic
Prevalence Study. These factors were weighted using logistic regression analyses to develop a clinical index that
was validated on 486 men and 484 women reported in the Sant Qubec cardiovascular health survey. A high
CHL/HDL ratio was defined as $5 for women and $6 for men which approximates the 75th percentiles reported
in the second United States National Health and Nutrition Survey. In the Lipid Research Clinics cohort, 307
men (15.4%) and 188 women (11.5%) had an elevated CHL/HDL ratio. Using separate clinical indices for men
and women, significant variables included body mass index, alcohol consumption, age, smoking status, systolic
blood pressure, physical activity status, and the presence of diabetes, the study identified 88% of the men and
82% of the women with elevated ratios. External validation using the Sant Qubec data set demonstrated test
sensitivities of 81% for men and 94% for women. Overall, 12% of those with a high CHL/HDL ratio were
misclassified as low risk. The ratio of total plasma cholesterol to HDL cholesterol has been shown to be one of the
best lipid predictors of increased coronary risk. Readily available clinical data can be used to identify 88% of
those individuals most likely to benefit from lipid screening while obviating the need for such screening in one
quarter of otherwise healthy adults. J CLIN EPIDEMIOL 52;1:4955, 1999. 1999 Elsevier Science Inc.
KEY WORDS. Hyperlipidemia, clinical index, coronary disease, risk prediction

INTRODUCTION
Identifying and treating individuals with risk factors for
coronary heart disease (CHD), such as hypertension or hyperlipidemia, is increasingly recognized as an important
CHD prevention strategy. While the cost of measuring
blood pressure is negligible when performed during routine
office examinations, serum lipid measurements require substantial health service resources if all adults are to be tested
according to current expert guidelines [1,2].
Garber et al. projected that in the United States, by the
year 1995, the cost to screen over 33 million people aged 65
and older would amount to between 50 and 62.2 million
dollars annually [3]. Similarly, Grover et al. estimated that
the initial costs to implement a nationwide cholesterol
screening program among Canadians would be 432 to 560
million dollars [4].
Given the increasing concerns surrounding limited health
care dollars, it would be useful to identify a more selective
*Address for correspondence: Dr. Steven A. Grover, Centre for the Analysis of Cost-Effective Care, Montreal General Hospital, 1650 Cedar Avenue, Montreal, Quebec, H3G 1A4 Canada.
Accepted for publication on 1 September, 1998.

screening strategy that obviates the need for lipid testing

among those individuals who are unlikely to require intervention. To this end, we have developed a clinical algorithm for identifying individuals at low risk of significant
serum lipid abnormalities.
Among commonly available serum lipid measurements,
the ratio of cholesterol to high density lipoprotein cholesterol (CHL/HDL) has been found to be one of the most accurate predictors of future coronary events [510]. Accordingly, we have used recursive partitioning techniques on
two representative adult populations to develop and validate a simple clinical decision rule for identifying individuals most likely to have elevated CHL/HDL levels [11,12].
The results of these analyses demonstrate that it is feasible
to identify a substantial portion of the North American
population as being at low risk of an elevated CHL/HDL ratio thereby restricting lipid testing and reducing the cost of
screening.
METHODS
We used the Lipid Research Clinics (LRC) Program Prevalence Study Cohort to develop the prediction model. The

S. A. Grover et al.

50

methods and design of the LRC study have been reported

previously in detail [1316]. Specifically, we used the 15%
random subsample, including 2510 men and 2316 women
aged 3074, to support the generalizability of our results.
After deleting those with incomplete clinical and laboratory data (345 men and 524 women) there remained 2165
men and 1792 women. We excluded those who had diagnosed cardiovascular disease (CVD) including CHD, cerebrovascular disease, or peripheral vascular disease at entry.
Among those free of CVD, we also excluded individuals
taking prescribed medication for hyperlipidemia and pregnant individuals. This resulted in 1993 men (92.0%) and
1631 women (91.0%) being included in the analysis.
In the LRC study, all men and women were followed up
prospectively to provide data on subsequent mortality.
Telephone or mail contact began annually in July 1977,
and individuals were followed up through June 1987, for an
average follow-up of 12.4 years. Specific causes of mortality
were ascertained by death certificate and hospital records,
and the vital status of 99% of the participants was established at least once during the follow-up period [17]. Any
LRC-defined definite or suspected coronary death was included in the present analyses.
Recursive Partitioning
Men and women were analyzed separately. A high CHL/
HDL ratio was defined as $5.0 for women and $6.0 for
men which corresponds approximately to the 75th percentile in U.S. population reported in the second National
Health and Nutrition Survey (NHANES II) [18].
Beginning with the complete cohort (root node), the population was partitioned into two new populations (nodes)
based on a splitting rule which maximized the difference in
the proportion of cases with a high CHL/HDL ratio. Each
candidate variable in turn was examined at all possible cutpoints. That cutpoint which produced the best split was
then used to subdivide the root node into two new populations. Each new split created a path or branch from the parent node (the population that was partitioned) to its sibling
nodes (the populations created from the split). The newly
defined populations were then partitioned using the same
criteriahence the recursive process. The procedure continued until all subpopulations could no longer be subdivided according to a predefined stopping rule. A node or
subpopulation that could no longer be partitioned was labelled a terminal node. In our model we required all subpopulations to maintain at least 25 individuals.
Once the model had been completed, we eliminated any
spurious population splits by pruning. We pruned the tree
from the bottom up. At each node where the population
had been subdivided, we tested the null hypothesis that
there was no difference in high CHL/HDL prevalence between the two populations (P , 0.01). If the null hypothesis was satisfied the branches were recombined to form a

single population and the process was then repeated until

all nodes had been tested.
Finally, we combined different terminal nodes (subpopulations) of the tree into classes as defined by an a priori clinical criterion. In this analysis, we classified as low risk any
terminal node where the subpopulation had at least a threefold lower risk of an elevated CHL/HDL ratio relative to the
total population. All other nodes were classified as high risk.

Clinical Prediction Index

To develop a clinical index for predicting a low CHL/HDL
level, we used logistic regression analysis to determine the
relative weights associated with each predictor variable
used in the final model. Using a low CHL/HDL ratio as the
dependent variable, we forced all significant binary splits
into the model. An odds ratio for each factor was then calculated from its b coefficient in the logistic model. These
odds ratios provided weighted points for each of the risk
factors such that a risk index could be created by summing
up all the points present in any one individual.

Validation
To validate the recursive models and clinical indices developed from LRC data, we applied the final results to a second, independent sample of adults unrelated to the LRC
cohort, using the Sant Qubec data set [19]. This crosssectional survey of the Quebec population was undertaken
in 1990 to characterize the prevalence of various risk factors for CVD. Among 676 men and 716 women with complete data, 486 (72%) and 484 (68%), respectively, were eligible for model validation after excluding individuals with
CVD, those taking lipid lowering drugs, or pregnant women.

RESULTS
Factors Associated with a Low CHL/HDL Ratio
The coronary risk factors of the two populations were similar and are summarized in Table 1.
Using univariate analyses, we identified specific variables
that were significantly (P , 0.05) associated with a low
CHL/HDL ratio in the LRC cohort. For males, regularly
engaging in strenuous physical activity at least three times
per week, increasing alcohol consumption, being a nonsmoker, the absence of a family history of premature CHD,
and a body mass index (BMI) #26 kg/m2 were positively associated with a low CHL/HDL ratio ,6. Among females,
an increased alcohol consumption, non-smoking status, no
family history of premature CHD, the absence of diabetes,
pre-menopausal status, age #50, systolic blood pressure
,120, and BMI ,24 kg/m2 were positively associated with
a low CHL/HDL ratio ,5.

Identifying Adults at Low Risk for Significant Hyperlipidemia

51

TABLE 1. Population coronary risk factors

Risk factors
(n 5 1993)
Mean level (6SE)
Age (years)
Systolic blood
pressure (mm Hg)
Diastolic blood
pressure (mm Hg)
Body mass index
(kg/m2)
Total cholesterol
(mg/dL)
HDL cholesterol
(mg/dL)
Prevalence of (%)
Diabetes
Smokers

Lipid research clinics

Sant Qubec

Males
(n 5 1993)

Females
(n 5 1631)

Males
(n 5 486)

Females
(n 5 484)

45.3 (0.2)

46.1 (0.3)

44.8 (0.7)

45.1 (0.6)

124.9 (0.4)

119.8 (0.4)

126.3 (0.7)

119.7 (0.8)

81.0 (0.2)

76.5 (0.2)

78.3 (0.4)

74.2 (0.4)

26.3 (0.1)

24.2 (0.1)

25.6 (0.2)

24.3 (0.2)

206.3 (0.8)

203.6 (0.9)

205.0 (1.7)

200.3 (1.8)

46.6 (0.3)

60.5 (0.4)

46.9 (0.5)

55.7 (0.6)

3.6
37.0

1.8
33.1

4.5
30.7

4.6
30.4

Recursive Partitioning: Development Data Set

Among men, 15.4% (307/1993) had a CHL/HDL ratio $6.0
(Figure 1). A BMI of #26 kg/m2 was the strongest predictor
of a low CHL/HDL ratio. Other significant factors associated with a low ratio included not being a smoker, moderate alcohol consumption (at least one drink per week), and
being active (engaged in strenuous activity at least three
times per week). In total, the four low risk subgroups of men
(numbered 14 in Figure 1), comprised 22.9% (457/1993)
of the male cohort. Among this low-risk group, only 3.9%
(18/457) had an elevated CHL/HDL ratio $6 (Table 2).
We completed a similar analysis for 1631 women, among
whom 188 (11.5%) had a high CHL/HDL ratio $5.0 (Figure 2). A BMI ,24 was the best discriminator variable as it
identified 838 individuals among whom only 54 had an elevated ratio. Among those with a BMI $24, 134/793
(16.9%) had an elevated ratio. Within each of the two resultant populations, smoking status provided the next best
discriminator variable. The partitioning process continued
until we derived a subpopulation (numbered 1 in Figure 2)
with a very low prevalence (0/111) of elevated CHL/HDL.
These women had a BMI ,24, were non-smokers, were
#50 years of age, and consumed three or more alcoholic
beverages per week. Likewise, five other subgroups were
classified into the low-risk category, such that the combined low-risk subpopulations comprised 31.9% (520/1631)
of the female cohort. In this low-risk group, only 1.2% (6/
520) had a CHL/HDL ratio $5.0 (Table 2).
Model Validation
The final model performed well when applied to the Sant
Qubec data set as shown in Table 2. The prevalence of el-

evated CHL/HDL ratios were similar in both the LRC and

Sant Qubec data sets for males (15.4% vs. 13.8%) and females (11.5% vs. 14.5%). The prevalence of an elevated
CHL/HDL ratio among high-risk LRC and Sant Qubec
men was 18.8% versus 15.9% and for women was 16.4%
versus 18.3%, respectively. The low-risk groups demonstrated a similar low prevalence of an elevated CHL/HDL
ratio between study populations: 3.9% for men and 1.2%
for women in the LRC cohort versus 7.0% and 3.9%, respectively, in the Sant Qubec survey.
Clinical Index
Having selected variables with the recursive partitioning
model, we forced each valuable into a logistic regression
model where the outcome of interest was a high CHL/HDL
ratio. The final model included only variables that were independently associated with a high CHL/HDL ratio at a
significance level of ,0.05. Accordingly, some factors identified by the recursive model were ultimately dropped from
the logistic model (i.e., age for men).
Using the logistic regression odds ratios as weights (Tables 3 and 4), clinical indices for males and females were
created. For each index, points based on each factors association with an elevated CHL/HDL ratio were derived after
rounding the odds ratio to the nearest integer. Summing up
all the points for each individuals risk factors produces a final risk score. By inspecting the LRC results, we chose a total score of five or less to classify a woman as low risk and a
total score of two or less for men.
Using these simplified clinical indices, 532 (26.7%) men
and 736 (45.1%) women were classified as low risk using
the LRC data set (Table 5). In men, 7.1% (38/532) of the

S. A. Grover et al.

52

TABLE 2. Results of a recursive partitioning model to identify

adults with an elevated CHL/HDL ratio ($5 for women and

$6 for men)
Males
Risk

Overall
(Percent)
Lowc
(Percent)
Highd
(Percent)

Females

LRCa

SQb

LRC

SQ

307/1993
(15.4)
18/457
(3.9)
289/1536
(18.8)

67/486
(13.8)
8/114
(7.0)
59/372
(15.9)

188/1631
(11.5)
6/520
(1.2)
182/1111
(16.4)

70/484
(14.5)
5/129
(3.9)
65/355
(18.3)

5 Lipid Research Clinics Follow-up Cohort.

5 Sant Qubec Cardiovascular Health Survey.
c Defined by specific combinations of clinical factors as summarized in
text and Figures 1 and 2.
d Defined as all those who are not classified as low risk.
a LRC
b SQ

those with an elevated ratio as low risk in the Sant Qubec

survey.
CHD Mortality in the Lipid Research Clinics

FIGURE 1. The recursive model for men to identify those at low

risk of an elevated cholesterol/HDL ratio $6. Each oval node
represents a decision point where a specific factor separates
those at low risk (on the right) from those at high risk (on the
left). The square nodes represent terminal nodes 1 through 4
which summarize four combinations of variables that identify
low-risk men among whom the prevalence of an elevated cholesterol/HDL ratio is less than 5%. Significant factors include
body mass index (BMI), smoking status, engaging in strenuous
activity at least three times per week (active), age, the number
of alcoholic beverages consumed (drinks), and systolic blood
pressure (SBP).

low-risk group actually had elevated CHL/HDL ratios. Similarly, for women, the false-negative rate was 4.8% (35/736).
Overall, the clinical index misclassified 5.8% (73/1268) of
the LRC low-risk group.
To validate the clinical indices, we applied them to men
and women in the Sant Qubec data. The false-negative
rate among those with elevated CHL/HDL ratios was 7.7%
(12/155) for men and 2.1% (4/187) for women (Table 5).
Overall, the clinical index misclassified 4.7% (16/342) of

To further assess the sensitivity of this screening strategy,

we examined the number of CHD-related deaths that occurred over a follow-up period of 12.4 years among LRC
participants. Of the 520 women classified as being at low
risk for a CHL/HDL ratio ,5.0, no CHD deaths occurred
compared with 12 CHD deaths in the high-risk group (0%
vs. 1.08%; P , 0.05). Of the 457 men classified as low risk
for a CHL/HDL ratio ,6.0, only 2 (0.44%) CHD deaths
occurred, which was significantly less (P , 0.01) than the
42 (2.73%) deaths among 1536 men at higher risk of elevated CHL/HDL. These data further confirmed that the
risk of not screening low-risk individuals for hyperlipidemia
was extremely small in terms of CHD mortality. Moreover,
it demonstrates that lipid abnormalities tend to cluster
among patients with other risk factors such as obesity, hypertension, smoking, and a sedentary lifestyle. Accordingly,
the clinical index presented herein has identified those at
low risk of both lipid abnormalities and coronary events as
well. Among 56 adults who would eventually die a CHD
death, 54 (96%) would have undergone lipid screening due
to the presence of factors associated with both coronary disease and a high CHL/HDL ratio.
DISCUSSION
These analyses demonstrate the strong association between
lifestyle variables and the likelihood of an elevated CHL/
HDL ratio. In both men and women, increased CHL/HDL
ratios were positively associated with smoking and an elevated BMI and negatively associated with alcohol consumption. These results are consistent with previously reported
findings of the LRC cohort and other studies [2024]. On

Identifying Adults at Low Risk for Significant Hyperlipidemia

53

FIGURE 2. The recursive model for women to identify those at low risk of an elevated cholesterol/HDL ratio $5. Each oval node represents a decision point where a specific factor separates those at low risk (on the right) from those at high risk (on the left). The
square nodes represent terminal nodes 1 through 6 which summarize four combinations of variables that identify low-risk women
among whom the prevalence of an elevated cholesterol/HDL ratio is less than 5%. Significant factors include body mass index (BMI),
smoking status, age, the number of alcoholic beverages consumed (drinks), systolic blood pressure (SBP), and diastolic blood pressure (DBP).

the other hand, some factors were significant only for men,
such as physical activity; while age, diabetes, and systolic
blood pressure were independent risk factors only for
women. Although many of these lifestyle factors have been

previously associated with high total cholesterol or low

HDL levels, these gender differences in the present analysis
underscore the importance of separate screening criteria for
men and women [2527].

TABLE 3. Factors independently associated with an elevated CHL/HDL ratio $5 among men

Risk factor
Body mass index (kg/m2)
Alcohol consumption
,1 time per week
Smoking
Strenuous physical
activity at least
3 times per week

b 6 (SE)

P value

Odds ratio

Points

0.76 (60.13)

,0.0001

2.13

0.66 (60.14)
0.63 (60.13)

,0.0001
,0.0001

1.93
1.88

2
2

0.48 (60.16)

0.0023

1.62

S. A. Grover et al.

54

TABLE 4. Factors independently associated with an elevated CHL/HDL ratio $5 among women

b 6 (SE)

P value

Odds Ratio

Points

0.96 (60.18)
0.37 (60.18)

,0.0001
0.038

2.62
1.45

3
2

0.53 (60.28)
1.13 (60.27)
1.10 (60.17)
1.38 (60.43)
0.45 (60.17)

0.063
,0.0001
,0.0001
0.001
0.01

1.69
3.09
2.99
3.97
1.57

2
3
3
4
2

Risk factor
Body mass index (kg/m2)
Age .50
Alcohol consumption
14 times per week
,12 times per month
Smoking
Diabetes
Systolic BP .120

The final recursive model revealed that among those in

the LRC cohort without CVD, who were not pregnant or
receiving therapy for hyperlipidemia, 23% (457/1993) of
men and 32% (520/1631) of women were at low risk for an
elevated CHL/HDL ratio. Overall, this represents over one
quarter of this healthy population.
Similarly, in the Sant Qubec population used for
model validation, 23.5% (114/486) of healthy men and
26.7% of healthy women (129/484) were classified as low
risk. Overall, 25.1% of all Quebec adults aged 3074 would
have been identified as low risk for an elevated CHL/HDL
ratio thereby reducing the need for lipid testing.
Translating these multivariate analyses into useful diagnostic tools for clinical decision making requires a clinical
index that allows for the quick assessment of individual patients during the course of a routine office visit. The resulting indices are reasonably simple and require information
that is readily available based on a routine personal history
and very basic clinical exam.
For both males and females in the LRC data set, these
clinical indices provided excellent test sensitivity: 0.88 in
men and 0.82 in females. However, test specificity was
rather poor at 0.29 and 0.51 for men and women, respec-

TABLE 5. Results of two clinical indices to identify adults

with an elevated CHL/HDL ratio

Males
Risk
Overall
(Percent)
Lowc
(Percent)
Highd
(Percent)

Females

LRCa

SQb

LRC

SQ

307/1993
(15.4)
38/532
(7.1)
269/1461
(18.4)

67/486
(13.8)
12/155
(7.7)
55/331
(16.6)

188/1631
(11.5)
35/736
(4.8)
153/895
(17.1)

70/484
(14.5)
4/187
(2.1)
66/297
(22.2)

5 Lipid Research Clinics Follow-up Cohort.

5 Sant Qubec Cardiovascular Health Survey.
c Defined by specific combinations of clinical factors. For men, low risk
was defined as a score of #2 points using factors summarized in Table 3.
For women, a score of #5 points using factors summarized in Table 4 was
defined as low risk (see text and Figures 1 and 2 for details).
d Defined as all those who are not classified as low risk.
a LRC
b SQ

tively. This reflects our conservative approach, whereby we

tried to miss as few high CHL/HDL cases as possible. Nonetheless, these overall results are similar to other published
indices to predict perioperative cardiac events, ankle fractures, and postoperative delirium [2830].
A similar study has also been published by Kinlay et al. to
identify individuals with high levels of blood cholesterol
(.6.5 mmol/L) [31]. Independent risk factors included age,
a history of hypertension, and a history of heart attack. The
final model demonstrated a test sensitivity of 0.77 and specificity of 0.61. However, it is becoming increasingly clear
that an elevated cholesterol level alone is a poor predictor
of coronary risk compared to the elevated CHL/HDL ratio
used in our present study [510].
It should be noted that there remain a number of potential weaknesses with a clinical index based on LRC data
from the 1970s. Lipid levels have been changing during the
past 20 years and the LRC 15% random sample may not be
truly representative of the adult population. Furthermore,
the accuracy of patients self-reported behaviors may be less
than the results observed in a carefully conducted survey
further undermining the test sensitivity and specificity reported herein. Nonetheless, we note that independent validation of the clinical index on the Sant Qubec data set,
which is representative of the adult Quebec population in
the 1990s, addresses some of these concerns. Applying the
clinical index to this contemporary adult population, resulted in high test sensitivities of 0.81 for men and 0.94 for
women and low specificities of 0.34 and 0.44, respectively.
Overall, 12% of men and women with a high ratio were
misclassified as low risk.
We also recognize that primary prevention is increasingly focussing on identifying those individuals at high absolute risk due to multiple-risk factors rather than lipid abnormalities alone. A multifactorial risk approach may be
particularly attractive if it can be shown to further reduce
the false-negative rate observed in our analyses [6].
In conclusion, we have demonstrated that readily available clinical data can be used to identify a substantial group
of individuals at low risk of an elevated CHL/HDL ratio.
These individuals are also at very low risk of dying of coronary disease over the next decade. Incorporating this simple

Identifying Adults at Low Risk for Significant Hyperlipidemia

clinical indices into current lipid screening guidelines could

significantly reduce the number of lipid tests ordered among
otherwise healthy adults in our communities.
This work was supported in part by grants from Heath Canada and the
Dairy Farmers of Canada. Dr. Grover is a senior research scientist
(chercheur-boursier) supported by Le fonds de la recherche en sant du
Qubec. We thank Dr. Louise Pilote and Mr. Louis Coupal for their
helpful comments on earlier versions of this manuscript and Ms. Nadine
Bouchard for her excellent secretarial support.

55

15.

16.

17.

References
1. Expert Panel on Detection, Evaluation, and Treatment of
High Blood Cholesterol in Adults. Summary of the Second
Report of the National Cholesterol Education Program (NCEP)
Expert Panel on Detection, Evaluation, and Treatment of
High Blood Cholesterol in Adults (Adult Treatment Panel
II). JAMA 1993; 269: 30153023.
2. The Canadian Consensus Conference on Cholesterol: Final
Report: The Canadian Consensus Conference on the prevention of heart and vascular disease by altering serum cholesterol and lipoprotein risk factors. Can Med Assoc J 1988;
139(Suppl 11): 18.
3. Garber AM, Littenberg B, Sox HC. Cost and effectiveness of
cholesterol screening in the elderly. Health Program Office of
Technology Assessment. Washington, D.C.: U.S. Government Printing Office; 1989: 3540.
4. Grover SA, Coupal L, Fakhry R, Suissa S. Screening for hypercholesterolemia among Canadians: How much will it cost?
Can Med Assoc J 1991; 144(2): 161168.
5. Grover SA, Palmer CS, Coupal L. Serum lipid screening to
identify high-risk individuals for coronary death. Arch Intern
Med 1994; 154: 679684.
6. Grover SA, Coupal L, Hu X-P. Identifying adults at increased
risk of coronary disease: How well do the current cholesterol
guidelines work? JAMA 1995; 274(10): 801806.
7. Stampfer MJ, Sacks FM, Salvini S, Willett WC, Hennekens CH.
A prospective study of cholesterol, apolipoproteins, and the risk
of myocardial infarction. N Engl J Med 1991; 325(6): 373381.
8. Castelli WP, Anderson K. A population at risk: Prevalence of
high cholesterol levels in hypertensive patients in the Framingham Study. Am J Med 1986; 80(Suppl 2A): 2332.
9. Brunner D, Weisbort J, Meshulam N, et al. Relation of serum
total cholesterol and high-density lipoprotein cholesterol percentage to the incidence of definite coronary events: Twentyyear follow-up of the Donolo-Tel Aviv Prospective Coronary
Artery Disease Study. Am J Cardiol 1987; 59: 12711276.
10. Goldbourt G, Holtzman E, Neufeld HN. Total and high density lipoprotein cholesterol in the serum and risk of mortality:
Evidence of a threshold effect. Br Med J 1985; 290: 12391243.
11. Ciampi A, Negassa A, Lou Zihyi. Tree-structured prediction
for censored survival data and the Cox model. J Clin Epidemiol 1995; 48: 675689.
12. Goulet JR, Mackenzie T, Levinton CM, Hayslett JP, Ciampi
A, Esdaile JM. The longterm prognosis of lupus nephritis: The
impact of disease activity. J Rheumatol 1993; 20: 5965.
13. Central Patient Registry and Coordinating Center for the
Lipid Research Clinics. Reference Manual for the Lipid Research Clinics Prevalence Study, Vols. 1 and 2. Chapel Hill,
NC: Dept of Biostatistics, University of North Carolina; February 1974.
14. Lipid Research Clinics Program Epidemiology Committee.
Plasma lipid distributions in selected North American popula-

18.
19.

20.

21.

22.

23.

24.
25.
26.

27.
28.
29.
30.
31.

tions: The Lipid Research Clinics Prevalence Study. Circulation 1979; 60: 427439.
Heiss G, Tamir I, Davis CE, et al. Lipoprotein-cholesterol distributions in selected North American populations: The Lipid
Research Clinics Program Prevalence Study. Circulation
1980; 61: 302315.
Lipid Research Clinics Program. Manual of Laboratory Operations, Vol. 1: Lipid and Lipoprotein Analysis. Bethesda,
MD: National Heart, Lung, and Blood Institute, National Institutes of Health; 1974. US Dept of Health, Education, and
Welfare publication (NIH) 75-628.
Jacobs DR Jr, Melbane IL, Bangdiwala SI, Criqui MH, Tyroler
HA. Lipid Research Clinics Program. High density lipoprotein cholesterol as a predictor of cardiovascular disease mortality in men and women: The follow-up study of the Lipid
Research Clinics Prevalence Study. Am J Epidemiol 1990;
131: 3247.
Caroll MC, Sempos C, Briefel R, Gray S, Johnson C. Serum
lipids of adults 2074 years. United States, 197680. National
Centre for Health Statistics. Vital Health Stat 11(242). 1993.
Esrey KL, Coupal L, Grover SA. valuation du risque de maladies cardiovasculaires. Exemple dapplication dune mthode
multifactorielle. In: Sant Qubec: Et votre coeur, a va?
Rapport de lEnqute qubcoise sur la sant cardiovasculaire 1990. Montral, ministre de la Sant et des Services sociaux, gouvernement du Qubec; 1994: 95101.
Heiss G, Johnson NJ, Reiland S, Davis CE, Tyroler HA. The
epidemiology of plasma high-density lipoprotein cholesterol
levels: The Lipid Research Clinics Program Prevalence Study:
Summary. Circulation 1980; 62(Suppl IV): 116136.
Ernst N, Fisher M, Smith W, et al. The association of plasma
high-density lipoprotein cholesterol with dietary intake and
alcohol consumption: The Lipid Research Clinics Program
Prevalence Study. Circulation 1980; 62(Suppl IV): 4152.
Anderson KM, Wilson PW, Garrison RJ, Castelli WP. Longitudinal and secular trends in lipoprotein cholesterol measurements in a general population sample: The Framingham Offspring Study. Atherosclerosis 1987; 68: 5966.
Dwyer T, Calvert GD, Baghurst KI, Leitch DR. Diet, other
lifestyle factors and HDL cholesterol in a population of Australian male service recruits. Am J Epidemiol 1981; 114: 683
696.
Wolfe RN, Grundy SM. Influence of weight reduction on
plasma lipoproteins in obese patients. Arteriosclerosis 1983;
3: 160168.
Wood PD, Klein H, Lewis S, Haskell WL. The distribution of
plasma lipoproteins in middle aged male runners. Metabolism
1976; 25: 12491254.
Stein RA, Michielli DW, Glantz MD, et al. Effects of different
exercise training intensities on lipoprotein cholesterol fractions in healthy middle-aged men. Am Heart J 1990; 119(2):
277283.
Howard BV. Lipoprotein metabolism in diabetes mellitus. J
Lipid Res 1987; 28: 613628.
Goldman L, Caldera DL, Nussbaum SR, et al. Multifactorial
index of cardiac risk in noncardiac surgical procedures. N
Engl J Med 1977; 297(16): 845850.
Stiell IG, McKnight RD, Greenberg GH, et al. Implementation of the Ottawa Ankle Rules. JAMA 1994; 271(11): 827
832.
Marcantonio ER, Goldman L, Mangione CM, et al. A clinical
prediction rule for delirium after elective noncardiac surgery.
JAMA 1994; 271(2): 134139.
Kinlay S, Heller RF. Selective screening for high cholesterol
in Australian general practice: The Newcastle Cholesterol
Prediction Study. J Gen Int Med 1990; 5: 18.